Transcript Agile Archeology Part 1

Agile Archeology
1. Introduction
• Agile principles did not all come from a single source.
• Arose from many - each from specific experiences.
• They were experiences by their own works.
• Ideas arose from various ‘communities’ and
adherents.
• The four main areas of thought are:
–
–
–
–
1.
2.
3.
4.
Waterfall process
Iterative and Incremental Development (IID)
Agile Software Development, and
Lean Software Development
Introduction
• Many management ideas harken back to
William Demming and Jay Forrester.
• In fact, most management ideas came from
two distinct differentiating generations of
software engineering:
– 1. SDLC 1.0 – the Waterfall
– 2. SDLC 2.0 – the Iterative Method Wars.
• This last generation includes Agile.
Introduction
• Historically, we can view the evolution as
either being derived
– from issues faced in projects that led to the
careful articulation of methods or
– From personal preferences or philosophical /
cultural differences.
– Did ‘natural selection of approaches’ take place or
did we simply lose a valuable set of practices.
2. SDLC 1.0
2.1 Accidental Waterfall
• The Waterfall Method - not intended to be a big
deal nor a landmark publication
• Originated via Winston Royce’s landmark paper that
addressed
–
–
–
–
simplicity,
iterative development,
risk reduction, and the
accrual of knowledge from working software.
• His paper was misinterpreted, so some can assert
that Agile is the ‘correct interpretation of his paper.’
Comments
• To our author and others, it is amazing how strongly
‘traditionalists’ adhere to this process.
• According to lean technology thinking, these advocates refer
to traditionalists as Type II Muda – that is, waste necessary
due to the way the work is currently performed.
• Upon looking at the Waterfall process, we can see that it is
essentially a one-pass, serial, and batched process.
• This is a very slow and deliberate process – likely the ‘slowest’
where there is no concurrency.
• In truth, its only virtue is its simplicity and lack of coordination /
integration activities.
SDLC 1.0 Background
• Upon direction of the federal government in attempts
to develop better methods and best practices,
approaches to evolutionary and iterative practices
were of interest.
• Please note that DOD was very heavily involved with
Waterfall methodology for most of its contracts.
• The undersecretary for DOD admonished singlepass and batch processes
• See p. 23 for a copy of the letter.
U.S. Government & DOD - heavyweights
• Large contractors with big projects and huge
organizations developing software were at the basis
for Customer Collaboration over Contract
Negotiation, a fundamental value of Agile.
• But because everything in the government required
big contracts, Waterfall model was perpetuated with
all of its documentation.
• What resulted were huge Ghantt charts used for
tracking and predicting development, among other
things.
Another Heavyweight
• SDLC 2.0 arose from experiences of
developers combined with the (mis)
application of the waterfall method.
• A lot of research took place via Barry Boehm
at TRW Systems and research at USC.
• Barry’s methodology turned into the Spiral
Method, a variation of the waterfall method,
which bases continuation of development
primarily upon the risks encountered on each
‘cycle.’
• Observe the next figure
Spiral Methodology